Parsing JSON and CSV

Python ships with modules for both formats you need most: json for hierarchical data and csv for tabular data. Neither requires installation. Both follow the same basic pattern: pass in a string, get back a Python data structure.

Parsing JSON

json.loads() converts a JSON string into a Python object. Objects become dicts; arrays become lists:

import json

raw = '{"name": "Alice", "score": 42, "tags": ["fast", "reliable"]}'
data = json.loads(raw)

print(data["name"])      # Alice
print(data["tags"][0])   # fast
print(type(data))        # <class 'dict'>

json.dumps() goes the other direction — Python dict to JSON string. The indent=2 argument makes the output human-readable:

result = {"status": "done", "count": 7}
print(json.dumps(result, indent=2))

{
  "status": "done",
  "count": 7
}

Parsing CSV

csv.DictReader reads a CSV string (or file) and yields each row as a dictionary keyed by the column headers:

import csv
import io

raw = """name,score,status
Alice,42,done
Bob,35,pending
Carol,51,done"""

reader = csv.DictReader(io.StringIO(raw))
for row in reader:
    print(row)
# {'name': 'Alice', 'score': '42', 'status': 'done'}
# {'name': 'Bob',   'score': '35', 'status': 'pending'}
# {'name': 'Carol', 'score': '51', 'status': 'done'}

Notice that every value is a string — CSV has no type information. If you need score as an integer for arithmetic, convert it explicitly: int(row["score"]).

The same data in both formats

Here is the same three-record dataset expressed first as CSV, then as JSON. Both represent identical information; the format changes only the shape of the text:

CSV:

name,score,status
Alice,42,done
Bob,35,pending
Carol,51,done

JSON:

[
  {"name": "Alice", "score": 42, "status": "done"},
  {"name": "Bob",   "score": 35, "status": "pending"},
  {"name": "Carol", "score": 51, "status": "done"}
]

CSV is more compact for simple tables. JSON is unambiguous about types (42 is a number, not a string) and extends naturally if you later need to add nested fields.

Try it

Parse CSV, filter rows, and serialise the results as JSON:

This is the core loop of data wrangling: parse, filter/transform, serialise. The specific operations change; the loop stays the same.

Where to go next

Next: lab — wrangle — read a CSV of product inventory, filter by stock level, and write the filtered results to JSON.